MOSCOW, November 9The development of systems based on artificial intelligence (AI), which are used as a doctor’s assistant, requires the standardization of electronic medical records; structuring their management will speed up the training of AI and bring closer its widespread use in medicine, said the development director of Webiomed (a market participant in the National Technology Initiative » Healthnet») Alexander Gusev.
“»Currently, uniform standards are not used to encode clinical information, and interfaces for entering medical records often allow free entry by a physician without quality control and completeness. All this hinders the development of AI products,» he said at the National Technological Revolution conference 20.35″.
Gusev noted that the share of unstructured records in such medical records is about 80%. Typos, blank lines and gaps in texts also interfere with the creation of databases for training AI models. At the same time, he said, standardizing approaches to filling out electronic forms “could solve this problem.”
«At the same time, it is important to keep in mind the balance between ensuring patient privacy and technology development. It is important to regulate the conditions for access to biomedical data without violating the rights of the patient, and ensuring the safe use of anonymized medical data for developers and researchers,» — the expert summarized.
Before this, in early November, Gusev, in a conversation with, predicted that the digitalization of healthcare and the introduction of artificial intelligence will reduce the need for personal visits to doctors, so by 2030, in-person visits to clinics could be halved.
Свежие комментарии